Using biometric behavior to prevent users from inappropriately entering sensitive information

Abstract

Disclosed embodiments can pertain to utilizing biometric behavior to prevent inappropriate entry of sensitive information. Current biometric behavior of a user, specific to the user's interaction with an electronic form, can be acquired and compared to benchmark biometric behavior of the user associated with past interaction with the electronic form to create a comparison result. A similarity score based on the comparison result can then be established. It can be inferred that sensitive information was incorrectly entered into the electronic form when the similarity score satisfies a predetermined similarity threshold. The sensitive information can be identified, and a machine learning model can be triggered to generate a risk score based on the similarity score and the identified sensitive information. A user can be prompted to redact incorrectly entered sensitive information when the risk score satisfies a predetermined risk threshold.

What is claimed is: 
     
         1 . A method of protecting sensitive information, comprising:
 executing, on a processor, instructions that cause the processor to perform operations associated with protecting sensitive information, the operations comprising:
 capturing a current biometric behavior of a user that is specific to a user's interaction with an electronic form; 
 comparing the current biometric behavior to a benchmark biometric behavior of the user associated with past interactions with the electronic form; 
 using a machine learning model, determining a risk score based on the comparing; 
 determining a likelihood that sensitive information was mistakenly entered into the electronic form based on the risk score exceeding a predetermined threshold; 
 determining that the sensitive information was mistakenly entered into the electronic form based on the determined likelihood; and 
 in response to determining that the sensitive information was mistakenly entered, prompting the user with a highlighted correct location and instruction to redact the mistakenly entered sensitive information. 
   
     
     
         2 . The method of  claim 1 , the operations further comprising determining that sensitive information was mistakenly entered into the electronic form using natural language processing and the risk score. 
     
     
         3 . The method of  claim 1 , the operations further comprising prompting the user with a chatbot that provides instructions on how to enter the sensitive information correctly. 
     
     
         4 . The method of  claim 1 , the operations further comprising prompting the user with a text window that specifies instructions on how to enter the sensitive information correctly. 
     
     
         5 . The method of  claim 1 , the operations further comprising preventing the user from moving to a next form field until the sensitive information is entered correctly. 
     
     
         6 . The method of  claim 1 , the operations further comprising providing a light indicator responsive to the risk score. 
     
     
         7 . The method of  claim 1 , wherein the current biometric behavior represents a period of time the user takes to complete sections of the electronic form. 
     
     
         8 . The method of  claim 1 , the operations further comprising comparing the sensitive information to a known syntax of the sensitive information. 
     
     
         9 . The method of  claim 1 , further comprising creating an audible indicator responsive to the risk score. 
     
     
         10 . A sensitive information protection system, comprising:
 a processor coupled to a memory that includes instructions that, when executed by the processor, cause the processor to:
 capture a current biometric behavior of a user that is specific to a user's interaction with an electronic form, wherein the current biometric behavior includes a current mood, and wherein the electronic form is an electronic document including two or more field locations for entering data; 
 compare the current biometric behavior to a benchmark biometric behavior of the user associated with past interactions with the electronic form to create a comparison result; 
 using a machine learning model, determine a risk score in response to the comparison; 
 determine that sensitive information was mistakenly entered into the electronic form in response to the risk score exceeding a predetermined threshold; and 
 prompt the user with a message beside a correct field location, the correct field location including a correct location to enter the sensitive information, the message including instructions to redact the mistakenly entered sensitive information and to enter the information at the correct field location. 
   
     
     
         11 . The system of  claim 10 , wherein the instructions further cause the processor to train the machine learning model using the current biometric behavior and the benchmark biometric behavior. 
     
     
         12 . The system of  claim 10 , wherein the instructions further cause the processor to invoke the machine learning model to:
 trigger a convolutional neural network (CNN) to establish the risk score; and   train the CNN using the benchmark biometric behavior.   
     
     
         13 . The system of  claim 10 , wherein the instructions further cause the processor to display information on an electronic device explaining how to enter the sensitive information. 
     
     
         14 . The system of  claim 10 , wherein the instructions further cause the processor to display a chatbot on an electronic device to permit chat with the user to instruct the user how to enter the sensitive information. 
     
     
         15 . The system of  claim 10 , wherein the instructions further cause the processor to create an audible signal that describes to the user how to enter the sensitive information. 
     
     
         16 . The system of  claim 10 , wherein the instructions further cause the processor to prompt the user on how to enter the sensitive information with a text window. 
     
     
         17 . A computer-implemented method, comprising:
 capturing an action behavior of a user that is specific to a user's interaction with an electronic form document including two or more field locations for entering data;   comparing the action behavior to a benchmark behavior of the user associated with past interactions with the electronic form;   using a machine learning model, determining a risk score based on the comparing;   determining that sensitive information was mistakenly entered into the electronic form based on the risk score exceeding a predetermined threshold; and   in response to the risk score exceeding a predetermined threshold, preventing the user from moving on by inactivating a button to proceed until the mistakenly entered sensitive information is redacted and entered correctly in a correct location prompted to the user.   
     
     
         18 . The method of  claim 17 , further comprising collecting user action information that includes timing information related to how rapidly the user enters sensitive information. 
     
     
         19 . The method of  claim 17 , further comprising collecting user action information associated with entering sensitive information at a website.